Nov. 2014. Vol. 5. No.

06 ISSN2305-8269
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org

AN EXPERT SYSTEM FOR DIAGNOSING FAULTS IN MOTORCYCLE

OLANLOYE, DAUDA ODUNAYO

Lecturer, Emmanuel Alayande College of Education, PMB 1010, Oyo, Oyo State, Nigeria
E-mail: olanloyeo@yahoo.com

ABSTRACT

Most of the people especially in the Africa sub-region preferred using a 2-wheeled vehicle to a 4 –wheeled vehicle
personally or for commercial purpose because it is easily affordable and the cost of maintenance is very cheap. This
is so because of the poor state of economy in the region.
The principle of thermodynamics has proved that a functioning engine could at times develop one fault or the other
and there is therefore a need to diagnose such fault in order to repair or rectify it. This research work presented an
expert system for diagnosing faults and profound possible solution using artificial intelligence principles.
Different rules are defined using forward chaining and implemented with CLIP programming Language.

Keywords: Thermodynamics, expert system, Artificial Intelligence, Forward Chaining, CLIP

1. INTRODUCTION helpful in our movement, logistics and transport

One major difference between living and non living
things is that living creatures have the ability to
move from one place to another whereas non living
things cannot. Man is also a typical example of living
organism that moves from places to places for one
reason or the other especially when struggling for the
daily bread. This brings us to the idea of
transportation which is defined as the movement of
people, animals and goods from one location to
another.
For man, there are different types of transportation.
These include transportation by land, Water and Air.
Transportation by water is only possible where there
is a big river through which boats /ships can be used
as a means of transportation. Transportation by air is
considered too expensive and therefore not
affordable by average citizen especially in the Africa
sub-region. Therefore transportation by land seems
to be the most common type that is easily affordable.
For this purpose, commercial cars, buses, Lorries,
tankers are available for various transportation
services. For comfortability, individuals can also
afford to buy a car or bus or motorcycle for private
usage. Vehicles used in daily lives are no more
considered as a luxury but a necessity. They are very
activities [5].
Unfortunately, most of the average citizens
especially in the Africa sub-region cannot afford to
buy a 4-wheeled vehicle on their own because of
their poor state of economy. Therefore they prefer
buying a two-wheeled vehicle which is popularly
referred to as motorcycles. People prefer motorcycle
to a four wheeled transport for the following reasons:
It is very cheap and easily affordable by individuals.
It has been established that it is very economical in
terms of fuel consumption compared to 4-wheeled
vehicle. Again when you consider the issue of
maintenance, it appears to be very cheap to maintain
and the brand new spare parts are very much
available and quite affordable. No wonder, most
people at lower class can afford to have at least one
motorcycle nowadays either for personal or for
commercial purposes.
However, the subject of thermodynamics has proved
beyond all reasonable doubt that, there is currently
no such machine that can work continuously without
taking energy from outside [5]. Hence, there are
gradual wears and tears of the engine. The machine
therefore continues to develop one fault or the other.

1
Nov. 2014. Vol. 5. No. 06
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org
Motorcycle just like any other vehicle can develop
fault anywhere and anytime. Under such
circumstances, it might be extremely difficult to seek
for the service of mechanical or automobile
engineers. The service of roadside mechanic may not
also be readily available.
This research work is out to render a wonderful
assistance in such situation. The proposed expert
system for diagnosing motorcycle faults will also be
of good assistance not only to the owner or the riders
but also to engineers and roadside mechanics who
might have reason to seek for more knowledge or
assistance why they are on the field.
When motorcycle developed fault, the rider will have
to call for the service of an Automobile engineers or
a roadside mechanic simply because he/she has no
technical skill and knowledge required to diagnose
such faults. Even when the faults is not a major type,
the attention of an engineer or expert is still required
otherwise in an attempt to diagnose such fault, a
faulty diagnosis may be carried out which further
aggravate the problem an ground. Dependence on the
expert can be minimized if its expertise can be
documented into computer system [6][9].
This research work therefore proposes an expert
system for diagnosis of motorcycle faults. Expert
system is a branch of Artificial Intelligence which
gives the user the opportunity to interact with
computer to solve certain problem.
2. LITERATURE REVIEW
Different types of expert system have been
developed for various real life applications. A web-
based expert system for fish disease diagnosis has
been developed by Daoliang, Zetian and Yanqing
[3]. The system is being used in North China by fish
farmers. Similarly, Ahmad [1] developed an expert
system for car failure detection. The systems possess
150 rules to detect different types of failures and
their causes. It divides the system into 3-major parts.
These are start up state, run stable state, movement
state.
Salama et al [10] implemented an Expert Diagnostic
Assistance System for car failure and malfunction.
This system will be highly useful in assisting
mechanics for diagnosing vehicle faults. The system
consists of three (3) major parts. These include –
knowledge acquisition which capture knowledge
from the domain expert and retain such knowledge in
2
ISSN2305-8269
knowledge base, knowledge engineer, external
sources of data and system user.
The second part is Graphical user interface which
consist of problem solution, Application
programming, problem diagnosis. The third phase is
the system module which consists of reasoning
specification, inference engine knowledge base and
user advisor.
Kadarsah [4] developed a decision model for car
faults diagnosis where ES was used to assist car
owners, drivers and in experienced mechanics.
Milanović et al [2] also developed a motocultivator
fault diagnosis model using hybridization of ES and
DSS. Nana and Simonov [5] developed a mobile
vehicle expert system for automobile industry. The
system is made so simple so that vehicle owners and
drivers can detect their vehicle faults and problems.
Adsavakulchai [7] developed electric learning (e-
learning) expert system for car fault diagnosis using
19 rules of knowledge base collected from different
sources - books, journal, engineering website e.t.c.
Three knowledge base were used – car start
problem, break problem and cooling system problem.
Visual Basic and Microsoft Access were used in the
implementation of the system.
Nabende [6] in his M.sc thesis developed an expert
system for diagnosing heavy duty diesel engine
faults. The expert system uses Bayesian network
technology to represent faults and their related
causes and recommended repair actions.
2.1 ARTIFICIAL INTELLIGENT (AI)
METHODS
AI could be defined as a scientific method of making
machines such as computer capable of exhibiting
intelligent behaviour.
The term intelligence covers many cognitive skills
such as the ability to learn, understand, recognize,
and categorize in an attempt to solve real life
problems. Trend of development in the area of AI
today reveals that a lot of progress has been made in
the area of problem solving-concept and method for
developing a system that is capable of reasoning
about problem, to arrive at a meaningful solution
rather than calculating the solution.
Nov. 2014. Vol. 5. No. 06
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org
It comprises of different branches such as Expert
system, genetic algorithm, fuzzy logic, neural
network e.tc.
2.1.1 Artificial Neural Network (ANN)
Human brain consists of 100 billion closely
interconnected single processing elements known as
neurons. A simplified model of the neuron and their
operation gave birth to ANN. Series of data which
serve as inputs are used to train the network and
hence produce the appropriate solutions. With
newish data, the system is able to use its past
experience to solve the problem. If Training or
learning phase involved human intervention, it can
be described as supervised learning or else it is an
unsupervised learning.
They are very good at solving problem that are not
prone to algorithmic solutions e.g. pattern
recognition, decision support etc. It has the ability to
handle previously unseen, incomplete or corrupted
data.
2.1.2 Genetic Algorithm
This belongs to a field known as evolutionary
computation. Process of arriving at meaningful
solution includes:
(i) Survival of the fittest
(ii) Cross breeding
(iii) Mutation
In the process,
 A population of candidate solution is initialized
(the chromosomes)
 New generations of solution are then produced
making use of initial population. To produce
these solutions, selection, crossover and
mutation are used.
 Next generation are then produced from fitness
function which is used to evaluate the fitness of
the newly evaluated solution.
 The steps of generating solutions as well as the
evaluation continue until acceptable solution is
found.
2.1.3 Fuzzy System
Traditional logic is based on proposition. Any
proposition is either true or false. To solve real life
problem, there is always a need to make use of
3
ISSN2305-8269
partial proposition- partly true or partly false. In such
cases, imposing precision may be difficult and may
lead to less optimal solution. Situation of this nature
is better handled by fuzzy system which is able to
make effective use of imprecise information by
assigning degree of truth using fuzzy logic. It gives
the opportunity of expressing knowledge in vague
linguistic terms.
2.1.4 Expert System (ES)
Expert system therefore is one of the major areas of
AI that has to do with scientific method of making
machines to acquire human expert knowledge to
solve a particular problem in a given domain. Expert
system can explain why data is needed and how
conclusions were reached.
ES can be defined as computer program that combine
expert knowledge in a particular domain and
disseminate to others. An expert system is a program
that emulates the interaction a user might have with a
human expert to solve problem in a particular
domain [8].
The system continues to ask questions from the end
user and expect the end user to supply answers as
input by selecting one or more from options provided
by the system or by entering another set of data as
input. Such interaction will continue until the system
reaches the conclusion. The solution arrived at may
be an exact solution i.e. single solution or multiple
solutions arranged in logical order. The system will
equally explain the reason why it arrives at such
conclusion.
ES has ability to utilize incomplete or incorrect data.
In fact, giving only a partial data set, an expert is
likely to produce accurate result with high degree of
certainty in its conclusion. The degree of certainty
can be qualified in relative terms and concluded in
knowledge base. The certainty values are assigned by
the expert during the knowledge acquisition place of
developing a system.
2.1.4.1 Advantage of Expert System
 With expert system in place, the probability and
the frequency of making good decision is high.
This facilitates a sort of consistency in decision
making. The development of expert system to
solve different real life problem has made it
possible to distribute human expert.
Nov. 2014. Vol. 5. No. 06
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org
 In most cases, the development of expert system
will reduce the cost of decision making i.e. the
availability of ES make proper and effective use
of available data.
 It permits objectivity by weighing evidence
without bias and without regards to the user’s
personal and emotional reactions.
 It made it possible for human expert to have free
time and mind to concentrate on some other
meaningful activities.
 ES support modular structure. This thereafter
paves way for high degree of dynalism in solving
real life problems.
2.1.4.2 Expert System Methods
ES adopts different types of methods. Some of these
methods include:
Heuristic Reasoning: This is the type of method the
human expert will adopt in solving problems. It
could be referred to as rules thumb or expert
heuristics. This method allows the expert to arrive at
a good conclusion quickly and efficiently. Unlike
human expert, ES adopt symbolic manipulation with
heuristic inference procedure that is very close to
human thinking process. For the ES to adopt this
method, it makes use of the following approach:
Search Control: ES embark on searching in a
particular domain. Many techniques have been
employed for this purpose. This includes pruning,
branch and band, breadth-first search and so on.
Because of the importance of search process, it is
imperative to use good search control strategy in the
ES Inference process.
(1) Knowledge Base: different types of rules are
set during the development of an ES. The rules
are declarative representation of the expertise.
It is often expressed in the form of IF……..
4
ISSN2305-8269
Forward Chaining Method: In the development of
an ES, various rules are put in place. There is
therefore a need to check the condition part of the
rule to determine truth or false value of such rule. If
the condition is true, the action part of the rule is also
true.
This process will continue until a solution is arrived
at or a dead end is reached. This approach or method
is being referred to as data driven reasoning.
Backward Chaining: Unlike the forward chaining,
the backward chaining is used to backtrack from a
goal to the paths that lead to a goal. Hence, it could
be referred to as goal driven. It has been found
highly applicable when all outcomes are known and
not too large in term of size.
2.1.4.3 Components of Expert System
ES is made up of different components that interact
together for proper and efficient functionality of the
system. Some of the major components of ES
include knowledge base, inference engine, working
memory, knowledge engineer etc. the integration of
these components are illustrated in fig 1 below:
THEN. These set of rules form the knowledge
base of an ES
(2) Working Storage: one of the major aims of
developing an ES is to solve a particular
problem in a particular domain. The ES would
Nov. 2014. Vol. 5. No. 06
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org
have to be provided with specific or relevant
data to a problem being solved. Such data
constitute the working memory of the ES.
(3) Inference Engine: set of code in the ES which
is developed with central aim of drawing
inferences or recommendation from knowledge
base and problem specific data in the working
memory is referred to as inference engine.
(4) User Interface: ES user interacts with the
system. This would not have been possible if
there is no link between the system and the
user. The code that contains the dialogue
between the user and the ES constitute the user
interface.
(5) Domain Expert: the ES is meant to solve
problem in a particular domain. The main
experts are individuals who are expert in
solving the problem. They could be described
as human expert in the specific domain. The
experience/ knowledge of a domain expert will
be highly regarded in the development of an
ES.
(6) Knowledge Engineer: the individual that
encode the expert knowledge in the declarative
form (making use of set of rules) that can be
used by an ES in solving problem in the
domain can be described as Knowledge
Engineer.
(7) User: one of the major functions of an expert
system is to provide and make possible solution
available to the user who will be interacting
with the system to get advice/ solution.
3. STATEMENT OF THE PROBLEM
When a vehicle develops faults,
mechanical/automobile engineers are usually
employed for repair [5]. The gravity of the vehicle
fault may at times be minor and not so serious to
seek help from an automobile/mechanical engineer.
Just like other vehicles, a 2-wheeled vehicle called
motorcycle can also developed faults at any time and
hence the attention of a mechanic/engineer is needed
for repair. But the service of such engineer or
mechanic may not be readily available. Therefore
bicycle owners or riders need a system that will
provide immediate solution especially when the
degree of the fault is quite minimal.
4. OBJECTIVES OF THE STUDY
The objectives of this research work are:
5
ISSN2305-8269
1. To analyzed and design an expert system for
diagnosing faults.
2. To implement the system using clips
programming language
3. To improve the knowledge of motorcycle
owners and riders in terms of diagnosing
various faults in motorcycle. This will actually
minimize the maintenance cost of the
motorcycle.
4. To develop an expert system that can be used
as a tool to train inexperienced
mechanical/automobile/motorcycle engineers
in the area of faults diagnosis and repairs in
motorcycle.
5. SCOPE AND LIMITATION OF THE
RESEARCH
Some faults in motorcycle are major and cannot
therefore be solved by the rider or the owner and
hence a mechanical/automobile engineer is
consulted. This expert system will serve as a tool to
guide the engineers during the diagnosis and repair
process. Situations where the fault developed is a
minor type, the system will guide the rider or the
owner to carry out such basic repair. So, the
developed expert system is mostly concerned with
minor faults diagnosis and repair. The system will
also be found useful to diagnose some common
major faults and add to the knowledge of mechanical
/automobile engineer in the field. The faults
considered by this system are limited into 3 major
parts. These are:
1. Faults related to starting of the engine
2. Faults developed while the motorcycle is on
motion
3. Electrical faults
6. METHODOLOGY
To develop this system, the following steps were
carried out
1. Identification of the possible faults in
motorcycle.
2. Relevant information was obtained from series
of literatures, experienced mechanics and
engineers were also consulted on how to
diagnose and repair such faults.
3. The system was designed to categorize the
faults into 3 major classes which include:
startup, on motion and electrical faults as
shown in fig 2.
Nov. 2014. Vol. 5. No. 06 ISSN2305-8269
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org

4. From the information collected, different rules

were generated to form the knowledge base of
the system as represented in fig 3 and 4.
5. The system was implemented using clip

Fig 2: main menu of the system
;;;===================================
;;; Expert System For Motorcycle Troubleshooting
;;; This expert system diagnoses some sample problems with a
motorcycle
;;;===================================
;;;********************************
;;;* STARTUP RULES *
;;;********************************
(Defrule Rule1 "Engine Does not Turn-over"
(Selection A1)
(engine turnover no)
(starter turnover no)
(battery ok)
(positive jumper cable starter cable connected)
(big spark turns starter)
(engine turnover no)
=>(printout t "Problem: starter solenoid is Bad")
(printout t "Solution: Replace starter solenoid))
;;;************************************

Fig3: CLIPS representation of a Rule

Rule1: Engine Does not turn over

IF the Selection is A1 "Run Start up Rules"
AND the engine does not turn over
AND the starter does not turn over
AND the battery is fully charged
AND the battery jumper cable is connected with starter cable
AND there is heavy spark that turns the starter
THEN the starter solenoid is damaged.
Therefore, replace the damaged starter solenoid

Fig4: English representation of the above rule

6
Nov. 2014. Vol. 5. No. 06
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org
SAMPLE RULES FOR KNOWLEDGE BASE
(A) START UP
1. IF the motorcycle rider kick the engine over or lit
the starter button AND the engine does not turn over
THEN check the starter solenoid.
2. IF the starter solenoid clicks and the motorcycle
rider kicks the engine over or hits the starter button
AND the engine still does not turn over THEN the
check the charge of the battery..
3. IF the battery is fully charged AND the engine
still do not turn over THEN check the jumper cable,
connect the positive jumper cable to the starter cable.
4. IF there is a big spark with the connection in 3,
and starter turns AND the engine still does not turn
over THEN the solenoid is bad and should be
replaced.
5. IF the solenoid has been replaced AND the engine
still does not turn over THEN the starter cable is
broken and should be replaced.
6. IF the starter cable has been replaced AND the
engine still does not turn over THEN the starter is
bad and should be replaced.
7. IF the starter has been replaced AND the engine
still does not turn over THEN connect the positive
jumper cable to the starter motor
8. IF there is a big spark with the connection in 7,
and the starter turns THEN the starter gear is bad,
replace the starter gear
9. IF the starter gear has been replaced AND the
engine still do not turn over THEN the starter clutch
is bad and should be replaced
10. IF the starter clutch has been replaced AND the
engine still does not turn over THEN the engine is
struck. Seek for the assistance of an experienced
mechanical or automobile engineer with reports from
rule 1 to 10
11. IF the motorcycle rider turns the key or hit the
start button to start the engine AND the engine is
cranking slowly making it impossible for the engine
to start THEN check the battery probably it is weak
7
ISSN2305-8269
12. IF the battery is charged AND the engine is still
cranking slowly when ignited THEN replace the
battery with a new one.
13. IF the battery has been replaced AND the engine
is still cranking slowly THEN seek for assistance of
assistance of an experienced mechanical or
automobile engineer with reports from rule 11 to 12.
(B) MOTION
14. IF the motorcycle is shaking or jerking while on
motion THEN the spark plugs are bad and should be
replaced.
15. IF the spark plugs have been replaced AND the
motorcycle is still shaking or jerking while on
motion THEN check if there is free flow of fuel
from the carburetor to the engine.
16. IF there is no leakage of fuel from carburetor to
the engine AND the motorcycle is still jerking or
snaking on motion THEN check for air leakage
between the carburetor and the in-let manifold.
17. IF there is no air leakage air leakage between the
carburetor and the in-let manifold AND the
motorcycle is still jerking or snaking on motion
THEN seek for assistance of assistance of an
experienced mechanical or automobile engineer with
reports from rule 14 to 16.
18. IF the rider apply the clutch AND the clutch will
not disengage THEN the clutch plates has struck
together and should be loosed.
19. IF the clutch plates are struck and making it
impossible for the clutch to disengage THEN the
loose the plate by running in first or second gear,
pulling-in the clutch and locking up the rear brake.
20. IF rule 18 and 19 cannot free the clutch plate
THEN consult Mechanical or automobile engineer to
take the clutch apart and oil the plate.
(C) ELECTRICAL
21. IF the rider turns off the engine AND the
headlight is still on THEN there is an electrical
problem, battery terminal should be removed.
22. IF the motorcycle rider fixes the battery terminal
back, switches off the headlight AND the light is still
on THEN the fault is not a minor type, an expert
should be consulted with report 21 and 22
Nov. 2014. Vol. 5. No. 06 ISSN2305-8269
International Journal of Engineering and Applied Sciences
© 2012 - 2014 EAAS & ARF. All rights reserved
www.eaas-journal.org

CONCLUSION Graduate Studies in Partial Fulfillment for the

Award of Master of Science in Computer
Just like any other engine, a motorcycle engine can Science Degree of Makerere University.
also develop one fault or the other. The type of fault [7]. S. Adsavakulchai, N. Ngamdumrongkiat and E.
developed might not be a serious type and hence can Chuchirdkiatskul “E-Learning for Car Faulty
be handled by the rider or the owner. An expert Diagnosis”, International Journal of
system for diagnosing motorcycle fault has been Information and Communication Technology
presented in this research work to serve as a guiding Research, Vol. 1, No. 1, pp. 20-26, 2011.
tool to the owner or the rider especially when the [8]. S. Samy, Abu Naser, Abu Zaiter and A. Ola,
automobile or mechanical engineer is not readily “An Expert System For Diagnosing Eye
available. The automobile or mechanical engineer Diseases Using Clips”, Journal of Theoretical
will also find the system useful. and Applied Information Technology, 2008.
[9]. S. T. Deepa and S. G. Packiavathy “Expert
Though, the cost of maintaining a motorcycle is System for Car Troubleshooting”, International
cheaper compared to a 4-wheeled vehicle, the Journal For Research In Science & Advanced
developed system will further reduce the Technologies, Vol. 1, Iss. 1, pp. 46-49, 2012.
maintenance cost since the rider or the owner can [10]. Salama A. Mostafa1, Mohd Sharifuddin
carry out some of these activities with the assistance Ahmad1, Mazin Abed Mohammed1 and Omar
of the developed system. Ibrahim Obaid. Implementing an Expert
Diagnostic Assistance System for Car Failure
This will encourage more people to buy motorcycle and Malfunction. IJCSI International Journal of
for personal or for commercial use and hence Computer Science Issues, Vol. 9, Issue 2, No 2,
boosting the economy of the country. March 2012 ISSN (Online): 1694-0814.
http://www.ijcsi.org/papers/IJCSI-9-2-2-1-
REFERENCES 7.pdf
[11]. The Basics of Expert (Knowledge Based)
[1]. Ahmad T. Al-Taani. An Expert System for Car Systems. Copyright @ 1997 by JM &
Failure Diagnosis. World Academy of Science Co/AJRA
Engineering and Technology, 12(2005):4-7.
[2]. D. D. Milanović, M. Misita, D. Tadić and D. L.
Milanović, “The Design of Hybrid System for
Servicing Process Support in Small Businesses,
FME Transactions”, 2010, 38, 143-149.
[3]. Daoliang Lia, Zetian Fua, Yanqing Duanb
(2002). Fish-Expert: a webbased expert system
for fish disease diagnosis, Expert Systems with
Applications, 23, 311-320.
[4]. Kadarsah. S. E. Ricardo Nurzal, “A Decision
Support System for Car Fault Diagnosis Using
Expert System”, International Journal of
Information Sciences for Decision Making, N
2, 1998
[5]. Nana Yaw Asabere, Simonov Kusi-Sarpong.
(2012). Mves: A Mobile Vehicle Expert
System for the Automobile Industry.
International Journal of Engineering Research
and Applications (IJERA) ISSN: 2248-9622
Available at
http://www.ijera.com/papers/Vol2_issue6/F
H2611081123.pdf. Vol. 2, Issue 6, November-
December 2012, pp.1108-1123
[6]. P. Nabende (2006). An Expert System for
Diagnosing Heavy Duty Diesel Engine Faults.
A Project Report submitted to School of